leaked directx8.1 build 620

Page 2 - Seeking answers? Join the Tom's Hardware community: where nearly two million members share solutions and discuss the latest tech.
G

Guest

Guest
my lowest scores where with high polygon count tests (so my cpu was responsible for that)
my duron 900 is like your athlon at 800 more or less... so our systems are comparable.

that 14 fps min in scene 4 and 16 fps min in scene 10 shows some problem with geforce with many textures I hear that even geforce 3 have some problems with texture management like with ultima IX (getting worse score in that game than even kyro 1 or radeon LE) or maybe it is those bandwidth problems of geforce 2 MX ... (radeon and kyro excels in that).
when the video card is more stressed geforce 2 mx start giving unplayable performance unlike kyro...
in those stress situations MX is giving half the performance of kyro 2.

Try a game with many explosions and you will see a geforce 2 MX with difficulty.

I would like to see kyro 2 with this bench with an high end cpu...
and geforce 2 GTS scores also...
Volunteers ?
Noko I am looking for those radeon scores also ;)...


Noko I found something that is kinda suspicious:

geforce mx 400 athlon 800
scene 4 5 6
min 14 30 50
avg 23 52 70
max 84 132 84

summary for scenes (4,5,6) 2000 !!!


for my duron 900 and kyro 2
scene 4 5 6
min 31 57 33
avg 41 85 34
max 97 131 96

summary for scene (4,5,6) 1351 ???what???

stupid comclusion here... maybe the guy who made this bench think that scene 6 is more important or something (that is when the T&L of MX start delivering better performance in T&L than my duron 900), but look at scene 4 14 fps min geforce mx400 should deserve a lower score for that ...


<P ID="edit"><FONT SIZE=-1><EM>Edited by powervr2 on 06/10/01 11:25 AM.</EM></FONT></P>
 
G

Guest

Guest
I bet that geforce 2 gts will beat kyro 2 in this bench....
this must be an worst case scenario for kyro 2...
I don't see much overdraw here...

<P ID="edit"><FONT SIZE=-1><EM>Edited by powervr2 on 06/10/01 11:26 AM.</EM></FONT></P>
 

noko

Distinguished
Jan 22, 2001
2,414
1
19,785
Hmmmmm, interesting, all I did was copy and paste the data here so I don't know how they got the 2000 with the apparent lower numbers compared to your 1351. Well I will be doing my Radeon shortly, will let you know.

Well to eat your <b>C :smile: :smile: kie</b> and have it too, gotta get <b>Rade :smile: n II</b>
 

noko

Distinguished
Jan 22, 2001
2,414
1
19,785
Radeon 64 183/183, 1.2@1.33ghz T-Bird

2964 - XSMARKS Scene 1-12

1024x768
32bpp
6/10/2001

MIN 46 69 55 24 39 60 38 59 39 23 37 49
AVG 64 88 66 35 70 76 60 63 51 36 44 46
MAX 76 112 91 93 144 83 77 65 62 71 63 66

SUMMARIES 2048 2496 1649 3775

The Radeon had richer textures then the MX400, both where very good, Radeon water tests looked better. I might up the Radeon clock and retest. I wonder what the new Radeon64s would do since they are now clocked at 200/200. What do you think?


Well to eat your <b>C :smile: :smile: kie</b> and have it too, gotta get <b>Rade :smile: n II</b>
 

noko

Distinguished
Jan 22, 2001
2,414
1
19,785
Radeon again with new W2K 3118beta drivers, overclocked to 195/195, T-Bird 1.2@1.4ghz

3137 - XSMARKS Scene 1-12
1024x768
32bpp
6/11/2001

MIN 50 74 58 25 46 66 40 63 42 25 42 52
AVG 69 93 68 38 75 79 64 66 54 39 46 49
MAX 83 112 92 91 154 83 83 67 70 72 65 71
SUMMARIES 2217 2604 1735 3995

Tomorrow I will check out WinMe in this benchmark.<P ID="edit"><FONT SIZE=-1><EM>Edited by noko on 06/11/01 00:04 AM.</EM></FONT></P>
 
G

Guest

Guest
much better results for radeon...
what is the use for MX400 to get more than 100 fps if sometimes it get lower than 15 fps...when radeon or kyro allways deliver playable fps...

kyro 2 minimum fps are similar to radeon minimum fps...
that is a strong point for kyro 2 and radeon...

I still think that average fps don't tell the whole truth!
a video card could go to 400 fps and then drop to 5 and get 60 fps average.
other could stay in 60 ... so the same score.. but the later card is WAY BETTER !!!!

<P ID="edit"><FONT SIZE=-1><EM>Edited by powervr2 on 06/11/01 07:18 AM.</EM></FONT></P>
 
G

Guest

Guest
I still think that this bench give wrong conclusions ...
look here:

scene 10,11,12 radeon 195/195 tb at 1,4 ghz win2k
min 25 42 52
avg 39 46 49
max 72 65 71
score 3995

scene 10,11,12 kyro 2 175/175 with duron 900 win2k
min 29 56 61
avg 43 74 57
max 67 117 83
score 1608 ???

WHAT ???
BUAHAHAHA
it detects that is a kyro so the final score is divided by 4 ??? ... or it detects that is a radeon and then multiplies the score by 4 ???
;)
funny
<P ID="edit"><FONT SIZE=-1><EM>Edited by powervr2 on 06/11/01 08:26 AM.</EM></FONT></P>
 
G

Guest

Guest
noko what you think about this?
;)
that is a very interesting result ...
 

noko

Distinguished
Jan 22, 2001
2,414
1
19,785
I don't understand their summaries at all. I understand the FPS but how the total score figured I havn't a clue. Maybe I will spend a few moments and figure out there scoring. Yes your right, a bouncing around FPS no matter how high especially if it falls way below 30FPS is much worst then one that is consistent and rarely falls below 30 FPS. The Radeon was smooth throughout (remember this was on a 1.4ghz machine) while the MX400 had moments of slide show quality animation. I think if you had a faster processor especially with the Kyro2 it would just get better and better with this benchmark. The Radeon is at its limits with a 1.4ghz processor. Also remember this program wasn't tested on anything other then a nvidia GF card meaning it was optimized for that card only.

Well to eat your <b>C :smile: :smile: kie</b> and have it too, gotta get <b>Rade :smile: n II</b><P ID="edit"><FONT SIZE=-1><EM>Edited by noko on 06/11/01 01:50 PM.</EM></FONT></P>
 
G

Guest

Guest
yes!!!
he only tested it with nvidia cards...

that is a big plus point for the drivers of radeon and kyro 2 ..
without even being tested they run without any flaws!!!

If it was the other way around... maybe nvidia cards would not run at all ...

<P ID="edit"><FONT SIZE=-1><EM>Edited by powervr2 on 06/11/01 12:52 PM.</EM></FONT></P>
 
G

Guest

Guest
I send a email to bustard (the guy who made that benchmark) asking what was this problem and got an answer...


well i must say that on some scenes the most important data is minimum fps,
on some other it's medium fps on others it's maximum
i dont have the answer to this question, i dont think it's low i would also
need to know what processor moves this machine
and please bear in mind glexcess was developed on nvidia vid cards
anyway you can have a look at other people results here:
http://www.glexcess.com/xsmark
so that you can find out what's wrong with you system
all the best,
Paolo "Bustard" Martella
WebMaster at GLExcess


his answer was not conclusive I will try to ask again...
with some info from this topic..

<P ID="edit"><FONT SIZE=-1><EM>Edited by powervr2 on 06/12/01 11:42 AM.</EM></FONT></P>
 

noko

Distinguished
Jan 22, 2001
2,414
1
19,785
hmmmm, where the nVidia cards score the lowest MIN FPS he uses the MAX FPS for the benchmark, when they have a high MIN FPS he uses that. Not to objective when the MX had a few slide shows and yours didn't but yet the MX scores higher. I know, lets just use the MIN FPS as data and add up all the results:

487 Kyro2, 900mhz Duron - total for MIN FPS
419 MX400, 800mhz Athlon Classic - total for MIN FPS
538 Radeon, 1.33gh T-Bird - total for MIN FPS
. . . (kinda hard to compare with such a large cpu difference)

Well well, the Kyro2 smokes the MX400. Maybe we should take some 1600x1200x32 data, I predict the Kyro2 to be the best. Will test later.

Well to eat your <b>C :smile: :smile: kie</b> and have it too, gotta get <b>Rade :smile: n II</b><P ID="edit"><FONT SIZE=-1><EM>Edited by noko on 06/12/01 02:24 PM.</EM></FONT></P>
 
G

Guest

Guest
1431 - XSMARKS duron 900 kyro 2 (hercules 4500)
1600x1200
32bpp
6/12/2001
22:49
MIN 24 32 26 11 24 26 22 29 19 11 24 24
AVG 35 35 39 17 34 33 32 31 24 16 31 23
MAX 43 38 46 62 59 46 39 32 27 29 49 33
SUMMARIES 1109 1276 867 1647

look at those minimum fps not bad at all !!!
only 3 times it gets to unplayable results !!!
;)

so what we can judge from his results?
it's only in 1600x1200x32 that kyro 2 start being the bottleneck not the cpu... (I guess that there are some bandwidth issues with some tests, at 1600x1200x32 not bad at all)
 
G

Guest

Guest
well it looks that athlon 800 is better than my duron 900 for kyro 2...
in the forum of that benchmark there are another kyro user that got better results than myself with kyro 2 and an athlon 800...
 
G

Guest

Guest
noko please put your radeon and mx400 score at 1600x1200x32 please...
;)

<i>
":) no way pal, you misunderstood the meaning of summaries:
last summary (cpu tests - red color) is not referred to scenes 10,11,12 but
to scenes 2,7,8 (cpu tests - red color)
thanks a bunch for your interest and dedication, talk to ya later!

Paolo "Bustard" Martella
Webmaster at GL Excess "
</i>
now I understand your mighty athlon did that great score...
;)<P ID="edit"><FONT SIZE=-1><EM>Edited by powervr2 on 06/12/01 11:22 PM.</EM></FONT></P>
 

noko

Distinguished
Jan 22, 2001
2,414
1
19,785
Wow, I had to write a custom monitor device driver using PowerStrip3 beta to get this program to recognize 1600x1200x32 on my Radeon. Here are the results: (T-Bird at 1.4ghz, 183/183 Radeon clock, default.)

2224 - XSMARKS Scene 1-12

1600x1200
32bpp
6/13/2001
MIN 19 75 30 10 18 32 17 28 17 10 18 20
AVG 26 89 42 15 30 44 30 40 21 16 22 19
MAX 32 112 59 65 62 56 44 52 26 30 32 27
SUMMARIES 843 1451 809 3861

Will do the MX400 tomorrow, thanks for the clarification of the summaries.

Well to eat your <b>C :smile: :smile: kie</b> and have it too, gotta get <b>Rade :smile: n II</b>
 

noko

Distinguished
Jan 22, 2001
2,414
1
19,785
Kyro2 - 1600x1200x32, Duron 900mhz, Win9x?? W2K??
272 MIN FPS

Radeon - 1600x1200x32, T-Bird <b>1.4ghz</b>, W2K
294 MIN FPS

Not bad, not bad, Kyro2 beats the Radeon 9 out of 12 test at Min FPS even though the Radeon has a slightly higher benchmark. I am going to clock my T-Bird at 866 and retest to see effect of cpu on this benchmark also be able to compare the results to the Kyro2 more realistically.

Well to eat your <b>C :smile: :smile: kie</b> and have it too, gotta get <b>Rade :smile: n II</b>
 

noko

Distinguished
Jan 22, 2001
2,414
1
19,785
Interesting, T-Bird at 866mhz, redid 1600x1200x32, Radeon 64 default clock, W2K

1744 - XSMARKS Scene 1-12
1600x1200
32bpp
6/13/2001
MIN 19 47 31 10 18 32 17 26 17 10 22 20
AVG 26 61 42 16 31 44 30 38 21 15 23 19
MAX 32 77 61 64 62 55 44 43 26 30 44 27
SUMMARIES 843 1481 834 2547

Decreasing the CPU speed had a dramatic effect on the overall benchmark, the 1.4ghz T-bird speed was 28% faster at the same resolution. Lets take a look at Min FPS:

Kyro2, Duron 900, 1600x1200x32
272 Total MIN FPS

Radeon, T-Bird 866, 1600x1200x32, W2K
269 Total MIN FPS

Virtually a tie, or is it? The Kyro2 beats the Radeon once again 9 out of 12 times with a 3 difference lead. In this test the Radeon scales more on the top end then the MIN FPS. Interesting. PowerVR if you had a faster processor your Kyro2 would really smoke. Well I will do the MX400 later today.


Well to eat your <b>C :smile: :smile: kie</b> and have it too, gotta get <b>Rade :smile: n II</b>
 

noko

Distinguished
Jan 22, 2001
2,414
1
19,785
Interesting, PowerVR look at the two different Radeon results, CPU 1.4ghz and CPU 866mhz, there is only two or three test that is significantly higher causing the higher GLexcess mark. A number of tests at 866mhz scored higher then the same test with the CPU at 1.4ghz. What do you make of that?
<P ID="edit"><FONT SIZE=-1><EM>Edited by noko on 06/13/01 01:56 AM.</EM></FONT></P>
 

noko

Distinguished
Jan 22, 2001
2,414
1
19,785
Been learning Visual C++ and havn't had time to hook up the MX machine to my monitor. Sucks having one monitor. Maybe I shouldn't have given away my 17" monitor. Anyways as soon as I can I will benchmark the MX400, sorry.

Well to eat your <b>C :smile: :smile: kie</b> and have it too, gotta get <b>Rade :smile: n II</b>
 

noko

Distinguished
Jan 22, 2001
2,414
1
19,785
Well I got a 17" for my MX400 machine. It just doesn't do 1600x1200. Eventually I will rehook up my other monitor and do the test you requested. Here is a test on the MX400 in 16bit. MXs are 16 bit screamers and the quality is fairly good too. W2K, 800mhz Athlon classic.

3242 - XSMARKS Scene 1 - 12
1024x768
16bpp
6/16/2001
MIN 66 48 75 32 50 73 45 60 51 32 65 75
AVG 95 64 110 46 92 96 65 75 67 49 86 69
MAX 112 84 136 100 172 121 84 88 84 99 113 100
SUMMARIES 3008 3647 2109 2802


Well to eat your <b>C :smile: :smile: kie</b> and have it too, gotta get <b>Rade :smile: n II</b><P ID="edit"><FONT SIZE=-1><EM>Edited by noko on 06/16/01 12:42 PM.</EM></FONT></P>